Back to results
US 11,689,526 B2Granted
Ensemble method for face recognition deep learning models
AssigneePaypal, Inc.
InventorsXiaodong Yu, Quan Jin Ferdinand Tang, Jiyi Zhang, Yuzhen Zhuo
Life of this patent
Estimating the term…
Abstract
Aspects of the present disclosure involve systems, methods, devices, and the like for user identification using Artificial Intelligence, Machine Learning, and data analytics. In one embodiment, a verification system and method is introduced that can provide user authentication using parallel modeling for face identification. The verification system used includes a face identification module for use in the identification and verification using parallel processing of a received image with a claimed identity. The parallel processing includes an ensemble of machine learning models processed in parallel for optimal performance.
Claims20
A system comprising:
- a non-transitory memory storing instructions;
- and
- a processor configured to execute instructions to cause the system to perform operations comprising:
- receiving, via a wireless network communication, a request for a user identification, the request including user information and a user image;
- retrieving a claimed image corresponding to the user image based in part on the user information received;
- inputting the claimed image and the user image to a plurality of parallel learning models in a face identification module;
- determining, from the plurality of parallel learning models, a plurality of user image vectors from the user image and a plurality of claimed image vectors from the claimed image, wherein each of the plurality of user image vectors has a corresponding one of the plurality of claimed image vectors, and wherein the each of the plurality of user image vectors and the plurality of claimed image vectors have a distinct length of n-dimension features based on a corresponding one of the plurality of parallel learning models used for vector output;
- determining, by the face identification module, an initial score associated with each of the plurality of parallel learning models based on comparing each of the plurality of user image vectors to the corresponding one of the plurality of claimed image vectors;
- determining, by the face identification module, a final score incorporating the initial scores associated with each of the plurality of parallel learning models;
- and
- transmitting, over the wireless network communication, a user authentication in response to the request for the user identification when the final score meets an ensemble criteria.
8 more elements
A method comprising:
- receiving, via a wireless network communication by a face verification system, a request for a user identification, the request including user information and a user image;
- retrieving a claimed identity image corresponding to the user image based in part on the user information received;
- inputting the claimed identity image and the user image to a plurality of parallel learning models in the face verification system;
- determining, from the plurality of parallel learning models, a plurality of user image vectors from the user image and a plurality of claimed identity image vectors from the claimed identity image, wherein each of the plurality of user image vectors has a corresponding one of the plurality of claimed identity image vectors, and wherein the each of the plurality of user image vectors and the plurality of claimed identity image vectors have a distinct length of n-dimension features based on a corresponding one of the plurality of parallel learning models used for vector output;
- determining, by the face verification system, an initial score associated with each of the plurality of parallel learning models based on comparing each of the plurality of user image vectors to the corresponding one of the plurality of claimed identity image vectors;
- determining, by the face verification system, an ensemble score incorporating the initial scores associated with each of the plurality of parallel learning models;
- and
- transmitting, over the wireless network communication, a user authentication in response to the request for the user identification when the ensemble score meets an ensemble criteria.
5 more elements
A non-transitory machine-readable medium having stored thereon machine-readable instructions executable to cause a machine to perform operations comprising:
- receiving, via a wireless network communication, a request for a user identification, the request including user information and a user image;
- retrieving a claimed image corresponding to the user image based in part on the user information received;
- inputting the claimed image and the user image to a plurality of parallel learning models in a face identification module;
- determining, from the plurality of parallel learning models, a plurality of user image vectors from the user image and a plurality of claimed image vectors from the claimed image, wherein each of the plurality of user image vectors have a corresponding one of the plurality of claimed image vectors, and wherein the each of the plurality of user image vectors and the plurality of claimed image vectors have a distinct length of n-dimension features based on a corresponding one of the plurality of parallel learning models used for vector output;
- determining, by the face identification module, an initial score associated with each of the plurality of parallel learning models based on comparing each of the plurality of user image vectors to the corresponding one of the plurality of claimed image vectors;
- determining, by the face identification module, a final score incorporating the initial scores associated with each of the plurality of parallel learning models;
- and
- transmit, over the wireless network communication, a user authentication in response to the request for the user identification when the final score meets an ensemble criteria.
5 more elements
Dependencies are inferred from the claim text.